I don't see how this burden to search companies is a reason to weaken the rights of copyright holders.
Image thumbnails in search results would probably be covered under fair use no?
Frankly I'm surprised that copyright wasn't already enforced this way. Documents viewed on the world wide web are these ethereal things that are composed on the fly by client browser software as instructed by web server software. If the web server software instructs the client software to present a document which mixes non-copyright-infringing and copyright-infringing content, it seems eminently reasonable to me that this would be copyright infringement.
What if I distributed a bunch of mini printing presses that, when you pressed a button, produced a perfect copy of this year's best selling novel? Sure, I didn't actually distribute the novel, but I enabled a mechanism whereby the user, when using my device as intended, would end up with a copyright infringing document.
I think of the web browser in the same way. The servers tell it what to display. Therefore, if they tell it to display something that violates copyright, then the server has violated copyright.
Here's how I would make the rules if I could:
- Publishing web pages with links to copyrighted content where those links cause the display of the copyrighted content inline in the linking document, would violate copyright
- Publishing web pages with links to copyrighted content where those links do not cause the display of the copyrighted content inline in the linking document, but instead merely lead the user to the content, would not violate copyright
Analog: you can publish instructions on where to go to listen to a copyrighted song. You cannot publish a document which plays the copyrighted song to the user.